Search code examples
pythonnumpytensorflowobject-detection

Printing class associated with detected object

I am running the default iPython notebook from TensorFlow's ObjectDetection section: https://github.com/tensorflow/models/blob/master/research/object_detection/object_detection_tutorial.ipynb

I am able to print the coordinates of the annotation made by the model using the code below in the final cell of the notebook. 

with detection_graph.as_default():
  with tf.Session(graph=detection_graph) as sess:
    # Definite input and output Tensors for detection_graph
    image_tensor = detection_graph.get_tensor_by_name('image_tensor:0')
    # Each box represents a part of the image where a particular object was detected.
    detection_boxes = detection_graph.get_tensor_by_name('detection_boxes:0')
    # Each score represent how level of confidence for each of the objects.
    # Score is shown on the result image, together with the class label.
    detection_scores = detection_graph.get_tensor_by_name('detection_scores:0')
    detection_classes = detection_graph.get_tensor_by_name('detection_classes:0')
    num_detections = detection_graph.get_tensor_by_name('num_detections:0')
    for image_path in TEST_IMAGE_PATHS:
      image = Image.open(image_path)
      # the array based representation of the image will be used later in order to prepare the
      # result image with boxes and labels on it.
      image_np = load_image_into_numpy_array(image)
      # Expand dimensions since the model expects images to have shape: [1, None, None, 3]
      image_np_expanded = np.expand_dims(image_np, axis=0)
      # Actual detection.
      (boxes, scores, classes, num) = sess.run(
          [detection_boxes, detection_scores, detection_classes, num_detections],
          feed_dict={image_tensor: image_np_expanded})
      # Visualization of the results of a detection.
      vis_util.visualize_boxes_and_labels_on_image_array(
          image_np,
          np.squeeze(boxes),
          np.squeeze(classes).astype(np.int32),
          np.squeeze(scores),
          category_index,
          use_normalized_coordinates=True,
          line_thickness=8)
      plt.figure(figsize=IMAGE_SIZE)
      plt.imshow(image_np)
      s_boxes = boxes[scores > 0.5]
      height = 636
      width = 1024
      s_boxes[:,0] = s_boxes[:,0]*height
      s_boxes[:,2] = s_boxes[:,2]*height
      s_boxes[:,1] = s_boxes[:,1]*width
      s_boxes[:,3] = s_boxes[:,3]*width
      for s in s_boxes:
            print(s)
      break

The output I get:

output

I am trying to print the class associated with an annotation the model makes so the output should be something like the following (Given 'Dog' has index 1 in 'category_index'):

[  23.5806942    23.79684448  548.24536133  326.084198  ]: 1
[  63.68989563  401.32214355  609.81091309  996.93786621]: 1

OR 

[  23.5806942    23.79684448  548.24536133  326.084198  ]: Dog
[  63.68989563  401.32214355  609.81091309  996.93786621]: Dog

The main problem I am having is, that I cannot figure out how to index an element from 'classes' for a corresponding score > 0.5.

The visualize_boxes_and_labels_on_image_array function is here:

https://github.com/tensorflow/models/blob/master/research/object_detection/utils/visualization_utils.py#L323

Solution

  • classes can be indexed similar to boxes

    s_class = classes[scores > 0.5]
print(s_class)

    Will return [ 18. 18.] for the first example in the object detection iPynb. 18 Corresponds to Dog in the category_index